Re: [DISCUSS] Considering trying out uv for our CI workflows

Yep. It all looks good now and I re-ran last intermittently failing job: Final effect of it: * CI image (uncompressed) with uv is slightly smaller (3.5 GB vs. 3.9 GB) * regular code only PRs: same time to incrementally build image ~ 1m * adding/modifying dependency in the PR:: 12 m -> 6m : 50%

[RESULT][VOTE] Release Airflow 2.8.2 from 2.8.2rc3

Hello, Apache Airflow 2.8.2 (based on RC3) has been accepted. 4 "+1" binding votes received: - Ephraim Anierobi - Jarek Potiuk - Hussein Awala - Elad Kalif 4 "+1" non-binding votes received: - Phani Kumar - Rahul Vats - Scheffler Jens - Amogh Desai Vote thread:

Re: [DISCUSS] Considering trying out uv for our CI workflows

Thanks for the superb investigation and effort @Jarek Potiuk ! I quite like the performance improvement numbers uv brings in compared to pip. I see no reason not to switch to UV in prod images as well. I will take a look at the pull request soon. Thanks & Regards, Amogh Desai On Mon, Feb 26,

Re: [VOTE] Release Airflow 2.8.2 from 2.8.2rc3

+1 non binding Tested my contributions along with a few general DAGs. Things don't seem to have digressed. Thanks & Regards, Amogh Desai On Mon, Feb 26, 2024 at 1:57 AM Scheffler Jens (XC-AS/EAE-ADA-T) wrote: > +1 (non-binding) > Tested the PRs I contributed as fixes and can confirm all of

Re: [DISCUSS] Considering trying out uv for our CI workflows

I think I will get it green finally: https://github.com/apache/airflow/pull/37692. I know where the test flakiness was from. Generally speaking it turned out that there is no free lunch and - of course - cache from uv increased our CI image size significantly (by around 1.5G) - and it caused much

RE: [VOTE] Release Airflow 2.8.2 from 2.8.2rc3

+1 (non-binding) Tested the PRs I contributed as fixes and can confirm all of this is working: - #37672 Fix regression on trigger form error display on form validation - #37497 Allow pre-population of trigger form values via URL parameters - #37063 Bugfix Triggering DAG with parameters is

Re: [DISCUSS] Considering trying out uv for our CI workflows

Some more findings. Overall, I can confirm that with `uv` we will get significant - 60 - 70% on build image times. This will impact both CI but also `breeze` local rebuilds. I am getting closer to a mergeable state. I switched to https://github.com/apache/airflow/pull/37692 to test "upgrade to

Re: [VOTE] Release Airflow 2.8.2 from 2.8.2rc3

+1 (non-binding) Below is the list of items I have tested: - Tested regression DAG with Celery and k8s executor. - Tested issues 37497 372027 37254

Re: [VOTE] Release Airflow 2.8.2 from 2.8.2rc3

Small comment. I raise it here because it's been discussed in the ASF discussions about "advisors" (which is a new initiative in the Foundation that I am taking part of). It's really good if the (+1) is accompanied by "what did I do" - this is why I always try to explain what kind of testing I've

Re: [VOTE] Release Airflow 2.8.2 from 2.8.2rc3

+1 binding On Sun, Feb 25, 2024 at 12:03 PM Phani Kumar wrote: > +1 non-binding > > On Sun, Feb 25, 2024 at 4:00 AM Hussein Awala wrote: > > > +1 (binding) ran the same checks and tests as the RC2; all look good. > > > > On Sat, Feb 24, 2024 at 8:22 PM Jarek Potiuk wrote: > > > > > +1

Re: [VOTE] Release Airflow 2.8.2 from 2.8.2rc3

+1 non-binding On Sun, Feb 25, 2024 at 4:00 AM Hussein Awala wrote: > +1 (binding) ran the same checks and tests as the RC2; all look good. > > On Sat, Feb 24, 2024 at 8:22 PM Jarek Potiuk wrote: > > > +1 (binding): installed RC3 and ran a few DAGs. Tested reproducibility, > > licences,